Subj : Re: Memory visibility and MS Interlocked instructions
To   : comp.programming.threads
From : Joe Seigh
Date : Thu Sep 01 2005 08:49 am

Alexander Terekhov wrote:
> Joe Seigh wrote:
> [...]
> 
>>It's not part of the standard PC definition. 
> 
> 
> What standard?
>  
> 
>>                                             Not in the definitions
>>that I googled and not in the definition in the Andy Glew posting
>>and not in the definition in the ia32 manual.  
> 
> 
> Because:
> 
> http://citeseer.ist.psu.edu/ahamad92power.html
> 
> "We have recently learned [11] that the processor consistency..." 
> 
> [11] 1993-tr-68.pdf:
> 
> "Our definition of PC is different from the definition provided by 
>  Goodman [Goo91]; comparing these two definitions is difficult since 
>  the definition in [Goo91] is ambiguous (see [ABJ +93] for an 
>  alternative definition of Goodman's PC).
> 
> [ABJ +93] Go to Because.
> 
> Run this loop until Monday (or until you finally get it), Joe. Hth.
> 

I'm finding it impossible to argue with a moving target.  If I subtract
everything you say out, it pretty much sounds like ia32 loads are not
always guaranteed to be "in-order".  End of discussion.


-- 
Joe Seigh

When you get lemons, you make lemonade.
When you get hardware, you make software.

.